This summer CSULB Computer Engineering and Computer Science Professor Colleen van Lent received a NASA Faculty Fellowship award that brought her to NASA's Jet Propulsion Laboratory in Pasadena, where she is collaborating with JPL's Artificial Intelligence Group on software that will significantly improve NASA's ability to manage its Deep Space Network .
NASA's Deep Space Network consists of three large (up to 70 meters in diameter) networked antennas located in the Mojave Desert, Madrid Spain, and Canberra Australia. Moreover, these antennas are responsible for uplink and downlink communication with all deep-space spacecrafts. Some examples include Voyager, the Mars Exploration Rovers, and Cassini -- the spacecraft currently orbiting Saturn. As Dr. van Lent explains "Almost every major decision concerning the spacecraft is dictated by ground users, so there is a constant uplink of commands. For instance, if the Mars rover sees an interesting' rock, it asks a human whether or not it should investigate. If commands cannot be sent, time and resources can be wasted while the rover waits. In addition, the space craft has limited onboard memory, so if no antenna is available to receive the periodic downlink of data, the information will be lost."
NASA's collaboration with Dr. van Lent was brought in part by her expertise in the area of planning and reasoning capabilities of intelligent systems . As she explains, "What makes this project interesting is that you have limited resources and competing interests. For example, a spacecraft has a limited view period -- time that it can link to a particular antenna. So sometimes it must stop communicating with one antenna as it goes out of view, and switch to a new contact. Some missions require specific resources, for instance a 70 meter antenna rather than 26 meter. Competition for resources is always an issue."
"Currently the final step for resolving conflicts for midrange plans (looking ahead about six months) is still done by hand. An automated system creates a rough draft and then representatives of each mission meet to iron out conflicts. Our goal is to create a new interface that can utilize existing planning software. This will help minimize the number of conflicts in the first place and then also break deadlocks, facilitate a fair distribution of resources, and in general give the mission specialists more time to concentrate on other matters."
Dr. van Lent, who teaches both undergraduate and graduate AI courses within the CECS Department, believes her collaboration with NASA will not only enrich the content of these courses, but will also help CSULB students obtain future employment with NASA. "Currently there are more than a dozen opportunities for CSULB students to work at NASA and/or JPL", she explains. "These range from co-op positions, summer internships, to permanent part-time employment. Students who reside within 50 miles of JPL can work part-time while going to school. In fact, if they work 20 or more hours a week they qualify for benefits. In general, it is very difficult to get a permanent job with JPL after graduation unless you interned previously so getting an early start is a must. And they take on students from several majors, not just engineering and computer science." Please visit NASA's Education Gateway for more information on student employment opportunities.